`
`
`
`
`
`IN THE UNITED STATES PATENT AND TRADEMARK OFFICE
`
`In re Patent of:
`Stephen Barbour
`U.S. Patent No.:
`11/574,372 Attorney Docket No.: 54598-0001PS1
`Issue Date:
`February 7, 2023
`
`Appl. Serial No.:
`16/484,728
`
`Filing Date:
`August 8, 2019
`
`Title:
`BLOCKCHAIN MINE AT OIL OR GAS FACILITY
`
`
`
`
`SECOND DECLARATION OF VERNON KASDORF
`
`
`
`
`
`
`
`
`Exhibit 1036
`Crusoe v. Upstream
`PGR2023-00039
`
`1
`
`

`

`I, Vernon Kasdorf, declare as follows:
`
`I.
`
`ASSIGNMENT
`
`1.
`
`My experience and qualifications are set forth in my July 20, 2023 declaration
`
`(EX1004).
`
`2.
`
`I have been asked to provide additional technical expert opinions, including
`
`opinions concerning the arguments and evidence in Patent Owner’s Response.
`
`3.
`
`Crusoe has specifically asked for my analysis from the perspective of a POSITA in
`
`the bitcoin mining industry. To the extent this declaration provides opinions on subject matter
`
`related to the gas and oil industry, I am relying on the opinions of Dr. Michael Nikolau (EX1003
`
`and EX1035). To that end, I am relying on Dr. Nikolau’s review and analysis of Dickerson, Belady,
`
`Boot, and the MAGS System – which are all identified as prior art herein.
`
`4.
`
`In preparing this Declaration, I have considered all of the materials previously
`
`submitted by both Upstream (including Exhibits 2001-2023) and Crusoe (including Exhibits 1001-
`
`1038). I have also relied on my education, training, and experience, and my knowledge of pertinent
`
`literature in the field of the ’372 patent.
`
`II.
`
`CLAIM CONSTRUCTION
`
`a. “blockchain mining devices”
`
`5.
`
`Upstream argues that “mining device 12” is made up of many components,
`
`including “a controller 86, network equipment 88 such as a modem and a network switch, … and
`
`one or more mining processors 92 such as processors 92A-E.” POR, 4. Upstream also argues the
`
`claims require “blockchain mining devices” to include mining processors and a specialized
`
`network interface. POR, 4.
`
`6.
`
`However, the specification and claims only require that “mining device 12”
`
`contains a mining processor. EX1001, 15:21-22, 16:32-33, claim 1. There is no requirement for
`
`2
`
`

`

`the “blockchain mining device” to contain a network interface or any other components. You depo
`
`tr, 157:12-158:14 (“it may or may not be within the network mining devices”); 177:9-178:11.
`
`Indeed, while Dr. You endorsed “having the whole box consider as a blockchain mining device”
`
`(You depo tr, 152:15-22), he later opined that that the “blockchain mining device” is a subpart of
`
`Hash Generator (e.g., minus the generator, the container box, and/or the racks holding the miners).
`
`You depo tr, 160:8-162:21 (“some elements relevant to the function of the blockchain mining
`
`devices are part of this device”), 205:13-207:16 (“not everything in the hash generator would be a
`
`blockchain mining device”).
`
`b.
`
`“mining processor”
`
`7.
`
`Upstream argues “mining processor” somehow requires a “blockchain mining
`
`circuitry” and requires a certain processor speed or power, because “computation effort” is
`
`important. POR, 5-6.
`
`8.
`
`However, neither the specification nor the claim language requires the “mining
`
`processor” to include any specialized hardware in the form of a “blockchain mining circuitry.”
`
`EX1001, 17:9-15. In fact, the specification uses a “CPU” as an exemplary “mining circuitry,”
`
`confirming there is no special hardware requirement and a generic personal computer CPU would
`
`satisfy this term. You depo tr, 181:4-182:13 (“A personal computer could be adapted to meet all
`
`these specifications”). Indeed, even today, one can still mine crypto currency using a generic
`
`personal computer. EX1022, Title.
`
`III.
`
`A POSITA WOULD HAVE BEEN MOTIVATED TO COMBINE DICKERSON
`AND CRYPTOKUBE
`
`9.
`
`Both gas generators and blockchain mining data centers were well-known before
`
`the priority date. Barbour depo tr, 25:9-26:16 (“the associated gas stream is commonly used by oil
`
`3
`
`

`

`companies where necessary.”), 64:3-15. As discussed below, it is my opinion that a POSITA
`
`would have been motivated to combine these two.
`
`a. The motivation to combine is not based on hindsight
`
`10.
`
`Szhmigielski lists electricity cost as the first concern for industrial miners. EX1009,
`
`88 (“Industrial miners face almost the same issues as data centers: access to relatively cheap
`
`power…”), 90 (“One of the biggest costs for Bitcoin miners is the cost for electricity.”); EX2031,
`
`4 (“miners are rushing to Texas for cheaper electricity costs”). Thus, in search of cheap electricity,
`
`a POSITA would have been motivated to employ the cheap or free electricity generated from a gas
`
`generator. Petition, 20.
`
`11.
`
`Upstream argues that Dickerson and CryptoKube are from disparate fields and the
`
`combination is based on hindsight bias. POR, 9-11. However, people were thinking about such a
`
`combination way before the priority date. EX2022, 18:4-18 (“I spoke to multiple clients who –
`
`who wanted to place my CryptoKube on a natural gas generator that was attached to flare gas. This
`
`was in – between 2014 and 2016”); EX1016, 1 (“mining with free natural gas”); Barbour depo tr,
`
`246:18-247:11. In fact, Belady discloses using casing head gas to power a data center, and the ‘372
`
`discloses nothing more than a specific use of Belady’s data center—mining cryptocurrency.
`
`EX1011, Title (“gas supply shock absorber for datacenter power generation”) and [0004] (“natural
`
`gas is a byproduct of oil drilling”); You depo tr, 39:4-9 (admitting a data center can be used for
`
`cryptomining); EX1030, 1 (“Upstream Data Inc. specializes in the design and fabrication of Ohmm
`
`bitcoin mining datacenters”).
`
`b. Gas generators provide “reliable” energy
`
`12.
`
`Upstream alleges that a POSITA would not have been motivated to make the
`
`proposed combination because energy generated by a gas generator is not “reliable.” POR, 34-39.
`
`4
`
`

`

`13.
`
`However, in my opinion, “reliable” is a relative term and any power system can be
`
`modified to be reliable. You depo tr, 23:11-24 (“any power source…they could eventually be
`
`reliable”). For example, while wind and solar are not reliable by nature, they can be made reliable
`
`by including things such as a battery or a hydrogen storage. You depo tr, 22:11-25 (“they could be
`
`reliable”). The prior art explicitly teaches making power from a gas generator reliable by including
`
`a backup energy sources (e.g., the diesel generator in Dickerson and the gas supply shock absorber
`
`in Belady). EX1005, [0021] (“a stand-by diesel generator is provided”); EX1011, Abstract (“draw
`
`gas from storage during negative pressure spikes”).
`
`14.
`
`Upstream’s argument that power from a gas generator is “unreliable” is also wrong,
`
`considering the wide-spread use of such gas generators to power oil wells (Barbour depo tr, 25:9-
`
`26:16), and Upstream’s position that gas supply is “continuous.” POR, 7 (“the combustible gas is
`
`continuous (albeit at varying production rates)”). Dickerson’s system and the MAGS system power
`
`the oil production facility (EX1005, Abstract), and Belady’s system powers a data center (EX1011,
`
`Abstract), and there is no reason to think the energy supply is unreliable. In fact, one of Dickerson’s
`
`advantages is “high uptime.” Dickerson, [0032]. Dickerson’s backup diesel generator also ensures
`
`reliability. EX1005, [0021]. Belady further teaches improving reliability by using the gas supply
`
`shock absorber. EX1011, Abstract. Indeed, comparing to prior art, the ‘372 patent offers no
`
`inventive measures to make gas generators more “reliable.” While boasting the importance of the
`
`“optional controller,” Upstream admitted that in most cases, there is no need for the “optional
`
`controller” because the gas supply is abundant. Barbar depo tr, 216:17-217:5 (“they were just
`
`taking a small chunk”). In any case, a POSITA would have known how to set up the system to
`
`only use a “small chunk” of the gas supply to ensure reliability. Id.
`
`5
`
`

`

`15.
`
`Further, comparing to Belady’s traditional data center (e.g., a 911 call center),
`
`cryptomining does not need as much uptime. I have extensive experience in designing such data
`
`centers and it is my opinion that, comparing to a blockchain mining data center (e.g., CryptoKube),
`
`Belady’s traditional data center would require a more reliable power supply. An advantage of
`
`bitcoin mining is that you can pause and resume at will. You depo tr, 237:1-13 (“I don’t believe
`
`everything will be lost.”).
`
`IV.
`
`A POSITA WOULD HAVE HAD A REASONABLE EXPECTATION OF
`SUCCESS
`
`16.
`
`Upstream alleges the Petition did not allege reasonable expectation of success.
`
`POR, 39-41.
`
`17.
`
`It is my opinion that a POSITA would have had a reasonable expectation of success
`
`in combining Dickerson and CryptoKube because the combination requires no more than plugging
`
`an appliance to a power source. The Dickerson system is designed to provide power to power the
`
`oil production facility, and to export excess electricity to the grid. EX1005, [0039]. Further,
`
`Dickerson teaches that “[t]he modular design approach of the containerized solution allows for
`
`future design flexibility” and “[t]hrough the addition of future engine and membrane units,
`
`additional gas can be processed and the generation capacity can be increased.” Dickerson,
`
`[0034]. In the event that minor adjustments are needed (e.g., adjusting the amount of Spondoolies
`
`or adjusting the generation capacity), a POSITA would have had a reasonable expectation of
`
`success in making such adjustments.
`
`V.
`
`THE LIMITATION “CONNECTED TO THE NETWORK INTERFACE AND
`ADAPTED TO…COMMUNICATE WITH THE BLOCKCHAIN DATABASE” IS
`OBVIOUS
`
`18.
`
`Upstream alleges that the combination of Dickerson and CryptoKube does not
`
`disclose the “connected to… communicate” limitation of claim 1, because the Petition does not
`
`6
`
`

`

`identify a “mining processor” and the ASIC chip in the Spondoolies SP35 servers (“Spondoolies”)
`
`only handles the hashing function. POR, 41-50 (“mining was the only activity performed by ASIC
`
`chips”). Similarly, Upstream alleges that the combination of MAGS and Polivka does not disclose
`
`this limitation. POR, 62-65.
`
`19.
`
`However, the Petition states: “CryptoKube is built inside an ISO shipping container
`
`and is filled with spondooliestech, that is, blockchain mining devices, each necessarily having a
`
`mining processor” Petition, 23. Thus, the Petition identified the “mining processor” as a subpart
`
`of the Spondoolies and I have explained that this subpart can be an ASIC chip. EX2022, 59:9-
`
`60:16. While Upstream alleges the only function of the ASIC chip is “hashing” (solving math
`
`problems), in order to mine, the ASIC chip necessarily exchanges data (i.e., communicate) with
`
`the blockchain database. Barbour depo tr, 204:5-12 and 207:22-208:5 (“an ASIC chip needs to be
`
`controlled and have access to up-to-date information on the network.”); You depo tr, 34:6-9
`
`(“network connectivity is important and necessary”), 197:1-10 (“blockchain mining would require
`
`connections and communications with the blockchain database”), 199:23-200:1 (“exchanging with
`
`the blockchain database will be necessary”). Indeed, the term “communicate” requires nothing
`
`more than information exchange. You depo tr, 87:1-4. Thus, this limitation is necessarily met by
`
`the ASIC chip in the Spondoolies.
`
`20.
`
`Further, the claim language does not require a “direct” connection/communication,
`
`and allows for “indirect” connection/communication through other components (e.g., a controller).
`
`For example, the specification discloses methods of indirectly connecting the mining processor to
`
`the network interface through a modem and/or a network switch. EX1001, 15:36-40 and 16:61-67
`
`(connecting a mining processor to the antenna via a modem and/or a network switch). Just like
`
`other processors (e.g., CPU, GPU, FPGA), the ASIC chip in the Spondoolies is connected to the
`
`7
`
`

`

`network interface via other components (e.g., a controller, a cache, a bus, and RAM), and the claim
`
`language does not require the “mining processor” itself to perform the communication functions.
`
`21.
`
`Even under Upstream’s narrow view that claim 1 requires the “mining processor”
`
`to perform the communication functions, a combination of an ASIC chip and a controller would
`
`meet the “mining processor” limitation. Dr. You admitted that the “blockchain mining device” can
`
`be a subpart of Hash Generator (e.g., minus the generator, the container box, and/or the racks
`
`holding the miners) and the “mining processor” can be a subpart of the S9 miner (e.g., minus the
`
`front fan and/or the container box). You depo tr, 160:8-162:21 (“some elements relevant to the
`
`function of the blockchain mining devices are part of this device”), 205:13-207:16 (“not everything
`
`in the hash generator would be a blockchain mining device”), 211:23-213:19 (“this subset, if they
`
`meet all the claim language, would be considered as a mining processor”). By the same rationale,
`
`a subpart of the Spondoolies (e.g., the combination of the ASIC chip and the controller) can also
`
`be the “mining processor” that meets the “connected to… communicate” limitation.
`
`VI.
`
`UPSTREAM FAILS TO ESTABLISH THE REQUIRED NEXUS
`
`a. No presumption of nexus
`
`22.
`
`It is my opinion that Upstream’s products including the Hash Generator and the
`
`Hash Hut are not “coextensive” with the claims at issue. Upstream’s own patent US11659682
`
`covers the cooling path feature of Upstream’s products. EX1024 claim 1; Barbour depo tr, 132:11-
`
`133:6 (“maybe one to three dozen”). Another patent US11907029 covers the heat recirc function
`
`of Upstream’s products. EX1025, claim 1; (Barbour depo tr, 233:16-234:8). As shown below in
`
`Upstream’s marketing material, eight features (including heat recirc) were marketed as “why our
`
`loadcenters outperform the competition”:
`
`8
`
`

`

`EX1026, 2. By comparison, the claimed invention of the ‘372 patent (e.g., the combination of a
`
`gas generator and a blockchain data center) is found nowhere in Upstream’s marketing materials.
`
`Thus, Upstream’s products are not “coextensive” with the claims at issue, and the presumption of
`
`nexus does not apply.
`
`b. Commercial success
`
`23.
`
`Upstream asserts that the alleged “commercial success” of Uptream’s products
`
`support a finding of non-obviousness. POR, 19-22. However, Upstream failed to show nexus—
`
`that the claimed invention of the ‘372 patent is the reason of the alleged “commercial success.”
`
`24.
`
`Contrary to what Upstream alleges, Upstream’s products do not embody claims 1
`
`and 24 for various reasons. First, a significant amount of Upstream’s products are single device
`
`deployments (i.e., a system containing only a single Hash Generator or Hash hut) that do not meet
`
`claim 1 according to Upstream. You depo tr, 153:10-154:7 (“this photo does not meet the
`
`limitation”); POR, 22
`
`(“approximately 219 (~66%) were deployed
`
`in multi-system
`
`9
`
`

`

`configurations”); Barbour depo tr, 168:20-169:3. Second, an unknown amount of Upstream’s
`
`products are deployed to utilize power from either the electric grid or commercial sales gas lines,
`
`and do not meet claim 1 according to Upstream. You depo tr, 104:21-106:8 (“a commercial sales
`
`line, which is not a hydrocarbon production, storage, or processing facility”), 107:16-109:7 (“A
`
`power grid is different from a power generation facility”); Barbour depo tr., 96:15-98:2 (“probably
`
`does not embody claim 1”), 149:22-150:9 (“under a hundred Hash huts sold where the customer
`
`applied it to a utility connect”); 151:5-10 (“at least some customers”). As evidence of commercial
`
`success, Mr. Barbour counted 284 units sold from 2021 to the date of the POR. POR, 22. However,
`
`at least one third of these 284 units were single device deployments. POR, 22; Barbour depo tr,
`
`156:9-161:20, 168:20-169:3. And it is not clear how many of these 284 units were deployments
`
`drawing power from natural gas wells, commercial sales gas lines, or the electric grid. Barbour
`
`depo tr, 156:9-161:20 (“Any application that was pretty clear was an oil and gas application, I
`
`generally included.”).
`
`25. Moreover, Upstream’s website list eight features as “why our loadcenters
`
`outperform the competition.” EX1026, 2 and EX1027, 2. At least the “passive heat recirc for cold
`
`climates” feature was patented in Upstream’s patent US11907029. Barbour depo tr, 233:16-234:8.
`
`Thus, according to Upstream’s own website, the alleged “commercial success” is attributable to
`
`either non-patented features or features patented in other patents. Indeed, the claimed invention of
`
`the ‘372 patent (e.g., the combination of a gas generator and a blockchain data center) is found
`
`nowhere in Upstream’s marketing materials. In fact, an unknown amount of Upstream’s products
`
`were deployed without a gas generator. Barbour depo tr, 149:22-150:9 (“I would guess that we are
`
`under a hundred Hash huts sold where the customer applied it to a utility connect.”).
`
`10
`
`

`

`26.
`
`Further, the commercial success could be attributable to numerous other reasons
`
`such as marketing efforts, increasing electricity costs, increasing computational efforts required
`
`for bitcoin mining, and increasing bitcoin price. Indeed, the computation efforts required for
`
`bitcoin mining increases year-over-year and mining one bitcoin using the same device requires
`
`more and more electricity. You depo tr, 57:14-58:6. Also, while Upstream alleges the alleged
`
`“commercial success” was not driven by bitcoin price, its allegedly sales increased long with the
`
`bitcoin price surge in 2021. POR, 21-22 (explaining 1 unit sold in 2017, 2 units sold in 2018, 10
`
`units sold in 2019, 32 units sold in 2020, and 284 systems sold from 2021 to the date of the POR);
`
`(cite bitcoin price chart).
`
`c. Long-felt unmet need
`
`27.
`
`Upstream alleges that there was a long-felt unmet need in the industry regarding
`
`how to utilize casing head gas, and this need supports a finding of non-obviousness. POR, 22-25.
`
`However, Upstream failed to show nexus—that the claimed invention of the ‘372 patent solves
`
`the need to utilize casing head gas.
`
`28.
`
`In my opinion, there was no unmet need in utilizing casing head gas. It was common
`
`practice for oil and gas companies to employ on-site gas generators to utilize casing head gas—to
`
`power the oil pump or other things. Barbour depo tr, 25:9-26:16 (“the associated gas stream is
`
`commonly used by oil companies where necessary.”). Further, before the priority date, there were
`
`multiple other solutions to utilize casing head gas, including Dickerson, Belady, and MAGS. You
`
`depo tr, 244:4-250:24 (“the Mags system does describe some methods to process the gas”). In fact,
`
`Belady discloses using casing head gas to power a data center, and the ‘372 discloses nothing more
`
`than a specific use of Belady’s data center—mining cryptocurrency. EX1011, Abstract; You depo
`
`tr, 39:4-9 (admitting a data center can be used for cryptomining); EX1030, 1. Further, industrial
`
`bitcoin mining only gained popularity in the 2010s, after Satoshi Nakamoto mined the first bitcoin
`
`11
`
`

`

`block in 2009—it is only natural that it took some time for people to connect bitcoin miners to a
`
`gas generator.
`
`29.
`
`Further, the identified need is not tied to the claims at issue because the claims
`
`broadly cover systems that use natural gas produced as a primary product. Barbour depo tr, 162:14-
`
`24 (“any oil and gas application”); You depo tr, 103:11-104:6 (“I think it covers both”); Barbour
`
`depo tr, 146:10-17 (“It might.”). In fact, the claims may even cover systems drawing power from
`
`the electric grid. Barbour depo tr, 96:15-97:2 (“unless the electrical grid is comprised of a generator
`
`with source combustible gas, then perhaps”).
`
`d. Industry skepticism
`
`30.
`
`Similar to commercial success and unmet need, there is no nexus between the
`
`claimed invention and the identified industry skepticism, because the claims broadly cover systems
`
`that use natural gas produced as a primary product, and even gas from commercial sales gas lines
`
`or the electric grid.
`
`e.
`
`Industry praise
`
`31.
`
`As evidence of industry praise, Upstream cites to a K33 report that describes
`
`Upstream as a “leading company” that “pioneered the concept in 2017.” POR, 28-30.
`
`32.
`
`However, in my opinion, the alleged “industry praise” could be attributable to
`
`numerous reasons (other than the ‘372 patent), such as Upstream’s marketing efforts and market
`
`share—there is no nexus between the praise and the claimed invention.
`
`VII. THE LOAD CONTROL CLAIMS (CLAIMS 10-13, 35-37 AND 40) ARE
`OBVIOUS
`
`33.
`
`Upstream alleges that the ‘372 patent discloses solving the casing head gas problem
`
`by “modulating the mining power load.” POR, 4. However, this load control feature is found
`
`nowhere in the independent claims, and is not required in most practical implementations. Barbar
`
`12
`
`

`

`depo tr, 216:17-217:5 (“they were just taking a small chunk”). Moreover, this feature is obvious
`
`over Belady, which discloses adjusting the load based on available gas supply. Petition, 54-55;
`
`EX1011, Abstract (“If the gas storage has insufficient gas, and a negative pressure spike occurs,
`
`the data center throttles down or offloads processing.”).
`
`34.
`
`Regarding the load control claims (claims 10-13, 35-37 and 40), Upstream argues
`
`no motivation existed to modify CryptoKube to achieve load control, because, among other
`
`reasons, (1) the modification would involve significant changes, (2) a POSITA would have been
`
`motivated to shut off all mining servers considering the servers may be damaged by voltage and
`
`frequency drop in the power supply1, and cooling may fail when there is no power, and (3)
`
`CryptoKube differed from Belady’s datacenter in that CryptoKube cannot offload processing to
`
`other data centers; (4) CryptoKube’s DCIM cannot control individual ASIC chips within
`
`Spondoolies; and (5) no POSITA was knowledgeable in both disciplines (oil and gas and crypto
`
`mining). POR, 51-55.
`
`35.
`
`Upstream is wrong. First, in my opinion, a POSITA would have been motivated to
`
`modify CryptoKube with a reasonable expectation of success because CryptoKube was highly
`
`customizable, and optional DCIM software that could be programmed (1) to selectively shut off a
`
`subset of the Spondoolies based on temperature or power supply and (2) to ensure the cooling
`
`system is prioritized over the Spondoolies. Specifically, the DCIM software can be programmed
`
`to monitor the temperature of the data center in several different spots. As the temperature rises,
`
`the DCIM software communicates with the PLC that controls the VFD (speed control for the
`
`fans). The rising temperature would trigger the fans to increase their speed and remove more hot
`
`air. In extreme temperature situations, where the fans were at their top speed and the temperature
`
`1 There is no evidence of any “voltage and frequency drop” in either Dickerson or Belady.
`
`13
`
`

`

`was becoming critical, the DCIM software would start to shut down individual power outlets on
`
`the PDUs until the temperature comes back into acceptable range. The same can be said for gas
`
`supply monitoring. As the gas supply drops, the DCIM software would start to power down
`
`individual PDU outlets until the power draw is below the power supply and is operating in a
`
`manner safe for the miner Power Supply Units (PSUs). The fans would only be powered down
`
`when there is an electrical failure, or the amount of power available is less than the amount needed
`
`to run the fans. The fans would be the last devices to be powered down in the power reduction
`
`sequence.
`
`36.
`
`Second, the load control claims broadly cover both (1) shutting off a subset of
`
`Spondoolies and (2) shutting off all Spondoolies. Barbour transcript, 211:5-14 (“The mining
`
`activity can be adjusted, I guess, a number of ways. Obviously, well, cutting power to it is one”);
`
`You transcript, 225:15-23 (turning on and off meets the limitation), 228:16-25.
`
`37.
`
`Third, comparing to a traditional data center (e.g., a 911 call center), one advantage
`
`of bitcoin mining is you can pause and resume at will—there is no need to actively offload the
`
`processing from CryptoKube because the processing is automatically offloaded to other miners in
`
`the blockchain network.
`
`38.
`
`Fourth, the “increasing or decreasing” limitation in claim 10 requires nothing more
`
`than switching on or off the Spondoolies (e.g., all or a subset of Spondoolies). Barbour transcript,
`
`211:5-14; You transcript, 225:15-23 (turning on and off meets the limitation), 228:16-25.
`
`39.
`
`Finally, CryptoKube is highly customizable, and an oil and gas POSITA could have
`
`easily worked with the CryptoKube supplier to modify the DCIM software.
`
`VIII. SPECIFIC GROUNDS
`
`a. Ground 1 (Dickerson+CryptoKube)
`
`14
`
`

`

`40.
`
`In the POR, Upstream presents arguments regarding (1) hindsight, (2) secondary
`
`considerations, (3) motivation to combine, (4) reasonable expectation of success, (5) the
`
`“connected to… communicate” limitation. Most of these arguments have been addressed above.
`
`Infra Sections III-VI.
`
`41.
`
`Regarding Dickerson, Upstream also argues that Dickerson does not provide “free
`
`excess electricity” considering the equipment cost. POR, 31.
`
`42.
`
`However, in the proposed combination, Dickerson’s system is already implemented
`
`to power the oil well, and there is no additional equipment cost needed. Dickerson teaches that its
`
`system generates excess electricity and teaches selling excess electricity to the grid. EX1005,
`
`[0039]. When no grid is available, a POSITA would have viewed CryptoKube as a great solution.
`
`EX1005, [0039] and [0024]. Further, Dickerson teaches “The modular design approach of the
`
`containerized solution allows for future design flexibility” and “[t]hrough the addition of future
`
`engine and membrane units, additional gas can be processed and the generation capacity can be
`
`increased.” Dickerson, [0034]. Thus, to the extent that more power is needed, a POSITA would
`
`have been motivated to modify Dickerson as needed to increase the generation capacity, taking
`
`into consideration the equipment cost and the potential profit generated by CryptoKube. A
`
`POSITA would have done so with a reasonable expectation of success considering Dickerson’s
`
`modular design.
`
`b. Ground 2 (Dickerson+CryptoKube+Belady)
`
`43.
`
`Upstream repeats the Ground 1 arguments, which have been addressed above. Infra
`
`Sections III-VI. Regarding the load control claims, Upstream’s arguments also fail for reasons
`
`discussed above. Infra Section VII.
`
`c. Ground 3 (Dickerson+CryptoKube+Belady+Boot)
`
`15
`
`

`

`44.
`
`Upstream argues that the Petition did not identify with particularity, whether Boot’s
`
`engine or Dickerson’s engine is used. POR, 55-56.
`
`45.
`
`In Ground 3, the Petition referenced Ground 2’s Dickerson-CryptoKube-
`
`Belady989 combination. Petition, 67 (“A POSITA would have found it obvious to improve the
`
`Dickerson-CryptoKube-Belady-989 combination by further including Boot’s teachings including
`
`using Boot’s load bank to absorb excess power”). Thus, Ground 3’s combination uses Dickerson’s
`
`engine, adding Boot’s load bank.
`
`46.
`
`Upstream alleges that while Dickerson is a “full-facility generator,” Boot is a
`
`“pump-related generator,” and Dickerson uses a chiller, which eliminates the need for a loadbank.
`
`POR, 57-58.
`
`47.
`
`However, in my opinion, there is no evidence that Dickerson’s chiller can soak up
`
`all the excess electricity. In fact, instead of soaking up excess electricity, Dickerson teaches the
`
`chiller is included to produce NGL for economic and environmental purposes. EX1005, [0022].
`
`Similar to Boot, Dickerson’s power supply and demand may fluctuate and a POSITA would have
`
`been motivated to include Boot’s load bank to be safe.
`
`d. Ground 4 (MAGS+Polivka)
`
`48.
`
`Upstream largely repeats the Ground 1 arguments, which have been addressed
`
`above. Infra Sections III-VI. The same motivation to combine applies equally to MAGS and
`
`Polivka. Further, Upstream argues that, unlike Dickerson, MAGS does not produce excess power
`
`even occasionally, because MAGS separated input gases into three streams and all the energy
`
`produced by MAGS was already spoken for. POR, 59.
`
`49.
`
`However, in my opinion, similar to Dickerson, MAGS can power the oil production
`
`facility and can transmit excess electricity to the grid. EX1014, [0023] (“to power various oil field
`
`equipment 812, or transmitted to the grid 814.”), [0134] (“for either local use or sale to the grid.”),
`
`16
`
`

`

`[0148] and [0161]-[0162] (“sale to the grid”). Thus, similar to Dickerson, a POSITA would have
`
`understood there would be leftover power when there is no grid access. Petition, 19, 81-82.
`
`50.
`
`Upstream further argues that MAGS only produces 450kWh, which is not enough
`
`to satisfy Polivka’s requirement of 120.3 kW. POR, 60.
`
`51.
`
`However, in my opinion, the 450 kWh embodiment is just one example based on
`
`one MAGS-200. EX1014, [0144]. The output of MAGS can be adjusted when the gas supply is
`
`ample and multiple MAGS-200 can be used. Young teaches “Flexible capacity scaling with
`
`multiple deployments on a single site is made possible with machine-to-machine communication
`
`and prioritization.” Young, [0298]. Further, CryptoKube is customizable, and a POSITA can
`
`install as many Spondoolies as desirable. Vern depo Tr., 28:21-24. The same goes for Polivka. A
`
`POSITA would have been motivated to customize the system based on gas supply with a
`
`reasonable expectation of success.
`
`52.
`
`Upstream also argues in Young, the engine is rated for 150kW, but only “about
`
`100kW” is required and thus 50 kW would be left over for this engine. POR, 61.
`
`53.
`
`However, as discussed above, this disclosure is just one example based on one
`
`MAGS-200, and multiple MAGS-200 can be used. A POSITA would have been motivated to
`
`adjust the capacity of the system based on demand with a reasonable expectation of success,
`
`considering the modular design. Young, [0298] (“Flexible capacity scaling with multiple
`
`deployments on a single site is made possible with machine-to-machine communication and
`
`prioritization.”).
`
`54.
`
`Upstream also alleges that my previous declaration is contradictory because it
`
`identified the Spondoolies as both a mining device and mining processor. POR, 63-65.
`
`17
`
`

`

`55.
`
`There is nothing contradictory. The ‘372 patent uses the term “mining processor”
`
`broadly. EX1011, 3:16-17 (“The mining processor comprises a plurality of mining processors”).
`
`My previous declaration only listed the Spondoolies as an example of a “mining processor.” Other
`
`examples of “mining processor” include an ASIC chip within a Spondoolies, a combination of an
`
`ASIC chip and a controller, or other subparts of a Spondoolies.
`
`e. Ground 5 (MAGS+Plivka+Belady)
`
`56.
`
`Upstream argues that it is unclear how Belady’s gas shock absorber is combined,
`
`and there is no reasonable expectation of success because Belady-989 taught shutting off gas flow
`
`to the generator when supply gas pressure rises. POR, 65-67.
`
`57.
`
`However, in my opinion, Belady’s “pressure regulating valve” is just a safety
`
`feature, and under normal conditions, it should never be triggered, because excess gas can be
`
`diverted into storage during positive pressure spikes. EX1011, Abstract.
`
`58.
`
`Upstream repeats the Ground 1 arguments, which have been addressed above. Infra
`
`Sections III-VI. For the load control claims, Upstream argues there was no motivation to modify
`
`Polivka to achieve load control, because mining servers could be damaged by voltage and
`
`frequency drop in the power supply, and cooling may fail when there is no power. This argument
`
`has also been addressed above. Infra, Section VII. As discussed above, shutting off Spondoolies
`
`based on gas supply still meets the load control claims.
`
`f. Ground 6 (101)
`
`i.
`
`The challenged claims are extremely broad
`
`59.
`
`Upstream argues (1) the gas flow from hydrocarbon facility is continuous and
`
`cannot be turned off without consequences (“not a residential gas line”); (2) the crypto mining is
`
`energy intensive. POR, 70-73.
`
`18
`
`

`

`60.
`
`However, in my opinion, the challenged claims are extremely broad and cover gas
`
`produced as the primary product. Barbour depo tr, 162:14-24 (“any oil and gas application”); You
`
`depo tr, 103:11-104:6 (“I think it covers both”). It even covers gas from commercial sales lines or
`
`power from the electric grid. Barbour depo tr, 146:10-17 (“It might.”), 96:15-97:2 (“unless the
`
`electrical grid is comprised of a generator with