Note to reader: this post is the result of the editing of several articles as well as my own input.
- The dollar index has stabilized after implementation of the dollar bearish tariff scheme. The USD still provides traders with a yield premium
- Like in the second half of 2019, recent issues with bank reserve levels could take a sudden turn for the worse with little warning
- Everything has reached a ‘liquidity peak,’ the Fed will be forced to ‘capitulate,’ and Bitcoin is the first to detect signals of market rescue
- AI implementation is proving to be structurally deleterious to white collar job growth levels. The economy is losing higher paying jobs quickly
- The temporary upticks in tariff induced inflation is wearing off. Consumer inflation sentiment is dropping.
- Federal government interest outlays continue to mount and the Federal Reserve has to be cognizant of potential longer-term outcomes. The Fed must be receptive to any avenue that can prolong this timeline to reduce this expense.
Liquidity has peaked
Recently, there has been a significant divergence in the market’s expectations regarding the Federal Reserve’s interest rate path for December. Previously, due to moderate inflation and weak labor data, it was widely believed that another rate cut in December was almost certain. However, a series of hawkish statements from the Federal Reserve has tempered this optimistic outlook. Although some dovish signals were released by officials on Friday, the debate over the direction of year-end monetary policy remains unresolved.
Michael Hartnett, Chief Investment Strategist at Bank of America, stated in the latest ‘Flow Show’ report that the Federal Reserve is under pressure to continue cutting rates due to the impact of the current liquidity tightening across multiple asset classes, with the cryptocurrency market set to become the first indicator of any shift in central bank policy.
Hartnett pointed out that assets such as cryptocurrencies, credit, the dollar, and private equity have all shown signs of ‘liquidity peaks.’ Multiple rate cuts by global central banks over the past two years fueled speculative sentiment in the market, but recent hawkish remarks from the Federal Reserve have raised doubts about further easing policies in 2026. Cryptocurrencies have suffered heavy losses, with Bitcoin and Ethereum continuing to decline, highlighting the impact of liquidity tightening on risk assets.
Hartnett expects that the current weakness in U.S. bank stocks is signaling a pattern similar to December 2018, and the ongoing decline in liquidity-sensitive sectors could force the Federal Reserve to pivot towards an easing policy. Looking back at 2025, a cumulative total of 316 rate cuts by global central banks had created a liquidity boom, directly driving three major market phenomena: AI investment frenzy, sharp volatility in Japanese stocks, and cryptocurrency speculation.
Looking ahead to 2026, Hartnett predicts that the Federal Reserve will experience a repeat of ‘policy capitulation,’ being forced to initiate a rate-cutting cycle. Under this scenario, three types of assets are expected to benefit the most: First, long-duration zero-coupon bonds, whose duration advantage will directly capitalize on the valuation premium brought by declining interest rates; second, Bitcoin, the asset most sensitive to liquidity shifts, which historically begins to soar before full confirmation of bailout signals; and third, mid-cap stocks, which are highly sensitive to financing costs and will exhibit significant earnings elasticity and room for catch-up once rate cuts materialize.
AI is rapidly dragging down the white collar job market
‘AI Will Leave a Lot of White Collar People Behind’ and It’s Time to Shift to This Industry, According to Fortune 500 CEO
November 23rd
Ford (F) CEO Jim Farley warned earlier this year that AI could leave many white-collar workers behind and urged a renewed focus on trade skills.
Recent research from Microsoft (MSFT) suggests that manual labor jobs like roofers and tire builders are among the least likely to be replaced by AI.
Despite a declining percentage of people saying college is “very important,” households with a college degree still earn significantly more than those without one.
For the past year, CEOs and founders have sounded the alarm, warning that AI could result in job loss and dramatically change the workforce.
Ford CEO Jim Farley is one leader cautioning that white collar workers should prepare for the potentially transformative impact of AI, arguing that some should consider trade work instead.
“I believe that AI and new technologies have an asymmetric impact on our economy—a lot of things are helped a lot and a lot of things are hurt,” Farley said in an interview with Walter Isaacson, a journalist and author, at the Aspen Ideas Festival in June 2025.
“When you look at these openings in our economy, it’s very clear that the technology we’ve seen has left a lot of people behind. AI will leave a lot of white collar people behind.”
Japan’s Debt Crisis Intensifies Global Liquidity Concerns
Japan is facing a crisis of simultaneous bond and yen collapse. The 30-year government bond has fallen 5% over the past two weeks, with a total annual decline of 12%, marking its worst performance since the 1970s; the yen’s exchange rate against the dollar is approaching the 160 mark, hitting a 40-year low.
At the same time, there is a clear mismatch at the policy level. The newly appointed Japanese Prime Minister has introduced large-scale fiscal stimulus amounting to 3% of GDP, while the Bank of Japan’s actual policy rate remains at negative 2%. This combination of ‘expansionary fiscal policy + loose monetary policy’ has exacerbated the depreciation of the yen and the selling pressure on government bonds.
Japanese government bond yields have broken through key resistance levels, contrasting sharply with the recovery trend in global bond markets. Policymakers are caught in a dilemma: raising interest rates to control inflation could trigger a stock market crash, while maintaining an accommodative stance would keep the currency and government bonds under pressure. This crisis is spreading globally through the unwinding of carry trades, with rising Japanese bond yields potentially causing international capital to flow back, impacting dollar liquidity, and subsequently affecting the U.S. stock market, credit bonds, and the cryptocurrency market.
Signs of liquidity tightening are emerging across multiple markets, with the Federal Reserve’s policy shift imminent.
U.S. mid-cap stocks are currently exhibiting a significant divergence between valuation and performance. Despite trading at a low multiple of 15 times earnings and benefiting from easing trade frictions and the trend of manufacturing reshoring, their performance has remained under pressure throughout the year. This contradiction highlights that the Federal Reserve’s policy adjustments have significantly lagged behind the actual needs of the market.
The banking sector index fell below 140, and the brokerage index dropped below 950. These two sectors, which are highly sensitive to liquidity, have historically been leading indicators of policy shifts. As seen in the market experience of December 2018, they will be the first to react to potential liquidity easing.
Hartnett believes that signals indicating the need for the Federal Reserve to continue cutting interest rates have emerged. Over the past two years, global central banks have cumulatively cut interest rates 316 times, fueling ‘collective market euphoria.’ Investors generally expect more rate cuts by 2026, but the Fed’s recent hawkish statements have raised some concerns.
At the same time, Hartnett emphasized that when the Federal Reserve is forced to make significant interest rate cuts, the market will present numerous investment opportunities. Historical experience shows that a ‘surrender’ in central bank policies often leads to significant revaluation opportunities for risk assets.
Cryptocurrencies become a barometer of policy shifts.
Hartnett believes that digital assets such as Bitcoin will serve as the ‘canary in the coal mine’ for detecting changes in central bank policies, with their price movements providing important early warning signals for investors. This assessment is based on the cryptocurrency market’s high sensitivity and rapid responsiveness to changes in liquidity.
Despite recent heavy losses in cryptocurrencies, with both Bitcoin and Ethereum experiencing sharp declines, Hartnett believes that cryptocurrencies will be the first to detect the Federal Reserve’s market rescue actions. According to Bank of America’s November fund manager survey, cryptocurrencies account for only 0.4% of institutional asset allocation, but retail inflows into the cryptocurrency market reached a record $46 billion in 2025. Derivatives trading currently accounts for 74% of cryptocurrency trading volume, making it a frontier for liquidity and speculation.
While institutional investors’ allocation to cryptocurrencies remains limited, the substantial influx of retail funds reflects strong market expectations for liquidity easing. Once the Federal Reserve signals a policy shift, the cryptocurrency market may rebound first.
It might be apparent to everyone or not but with the latest release of Google Gemini 3 I am seeing a very similar trend in technology changes in the AI hardware space as was seen with crypto mining about 10 years earlier and so I used some ai models to create a list of the likely benefactors for potential investment choices. Feel free to evaluate on your own – I think some of the best moonshots are still private or foreign and more difficult to invest in.
MASTER AI HARDWARE WATCHLIST (2025)
Deduplicated, categorized, and scored for risk/reward asymmetry
1. CUSTOM AI ASIC & TRAINING-MEGACHIP PLAYERS
(Closest to “ASIC mining rigs” → replacing GPUs in hyperscalers and datacenters)
Company Ticker Status Role / Tech Crypto Rhyme Risk/Reward (1–10)
Broadcom AVGO Public Custom ASICs for Google TPUs, Meta MTIA, OpenAI ASICs (2026), packaging ASIC manufacturer for hire 7
Marvell MRVL Public Custom accelerators, AI networking silicon Mining rig interconnect builder 7
Alchip 3661.TW Public Custom ASIC design for hyperscalers, AI customers Antminer engineering team 8
Cerebras Systems CBRS IPO Watch Wafer-Scale Engine (WSE-3) for ultra-large model training/inference Wafer-scale ASIC miner 9
Groq Private IPO watch LPU deterministic inference chips Ultra-low-latency ASIC 10
SambaNova Systems Private IPO watch Reconfigurable Dataflow Architecture (RDA) Programmable ASIC miner 9
Tenstorrent Private IPO watch RISC-V vector engines + AI accelerator Custom RISC-V miner 9
d-Matrix Private IPO watch In-memory compute inference ASIC Radical accelerator design 10
Graphcore Private (recent setbacks) IPO unlikely soon IPU accelerator Niche ASIC miner 8
Category Winner Likelihood:
* Training: AVGO (now), Cerebras (mid-term), AWS Trainium/Google TPU ecosystem (long-term)
* Inference: Groq, d-Matrix, Qualcomm
2. EDGE AI / LOW-POWER ACCELERATORS
(Equivalent to “FPGA mining rigs → ASIC-lite edge devices”)
Company Ticker Status Role / Tech Crypto Rhyme Risk/Reward
Lattice Semiconductor LSCC Public Low-power FPGAs for edge inference FPGA miner 8
QuickLogic QUIK Public eFPGA IP for custom SoCs Embedded DIY FPGA 9
Ambarella AMBA Public CVflow vision SoCs Application-specific miner 7
Mobileye MBLY Public ADAS inference SoCs Automotive ASIC miner 6
Qualcomm QCOM Public On-device LLM inference (Snapdragon, Oryon) Mobile ASIC miner 6
NXP Semiconductors NXPI Public Industrial/automotive AI microcontrollers Industrial ASIC miner 5
Blaize BZAI Public (small-cap) Edge inference accelerators Low-power ASIC 9
Hailo Private − Edge inference ASIC Tiny ASIC miner 10
Category Winner Likelihood:
* LSCC + QCOM (volume)
* Hailo + Blaize + QUIK (asymmetry)
3. MEMORY, HBM, STORAGE, AND BANDWIDTH (THE REAL BOTTLENECK)
(Equivalent to “VRAM suppliers / cooling vendors” in crypto mining)
Company Ticker Status Tech Crypto Rhyme Risk/Reward
Micron MU Public HBM3E/HBM4 High-speed VRAM supplier 7
SK hynix 000660.KS Public (Korea) #1 HBM supplier Lifeblood of AI rigs 8
Samsung Electronics 005930.KS Public HBM + packaging + fabs Miner VRAM + hardware fab 6
Western Digital WDC Public Enterprise SSDs Cheap storage for inference 5
Kioxia Private JV with WDC High-bandwidth flash (HBF) Flash alternative to HBM 8
Rambus RMBS Public HBM/GDDR/PCIe/CXL controllers Memory controllers for ASICs 8
Category Winner Likelihood:
* SK hynix (HBM)
* Rambus (every custom ASIC needs their IP)
* Micron (US-based HBM)
4. ADVANCED PACKAGING, INTERCONNECT, SILICON PHOTONICS
(Equivalent to “mining farm builders + riser cables + power systems”)
Company Ticker Status Tech Crypto Rhyme Risk/Reward
Arista Networks ANET Public AI fabrics (Ethernet/400G/800G) Networking backbone 7
Cisco Systems CSCO Public Switching fabrics Networking backbone 5
Coherent COHR Public Silicon photonics, optical interconnects Next-gen optical miners 8
Entegris ENTG Public EUV/CoWoS materials Packaging for AI ASICs 8
Astera Labs ALAB Public PCIe/CXL/AI connectivity silicon “Motherboard” of AI clusters 9
Category Winner Likelihood:
* Astera Labs (ALAB)
* COHR (optical future)
5. FOUNDRIES & MANUFACTURING
(Equivalent to “Taiwan fabs minting the ASICs”)
Company Ticker Status Role Rhyme Risk/Reward
TSMC TSM Public Manufactures TPU, Trainium, Groq, Cerebras, AMD, Nvidia Fab of all miners 7
UMC UMC Public Lower-end ASIC fabrication Cheap miner manufacturer 6
GlobalFoundries GFS Public RF + automotive + edge fabs Volume miner producer 6
6. EDA / IP OWNERS (THE “GOLD MINE OF PICKS & SHOVELS”)
Company Ticker Status Role Crypto Rhyme Risk/Reward
Synopsys SNPS Public EDA tools + IP Pickaxe store 6
Cadence CDNS Public EDA + memory IP Same 6
ARM Holdings ARM Public CPU IP for AI SoCs Architect of miner boards 7
7. MEGACAP CLOUD “INTERNAL ASIC” PLAYS (SECOND ORDER EXPOSURE)
(These are not chip vendors but beneficiaries of GPU→ASIC shifts)
Company Ticker ASICs Built AI Role Risk/Reward
Alphabet GOOGL TPUv1–v7 (Ironwood), Trillium Massive TPU clusters; MoE-optimized 5
Amazon AMZN Inferentia, Trainium 50B+ inferences/day 5
Microsoft MSFT Maia, Cobalt Azure AI custom silicon 4
(Lower risk/reward because asymmetry is diluted in megacaps.)
8. NEUROMORPHIC / EXPERIMENTAL
(The “quantum miner” category — huge upside, huge uncertainty)
Company Ticker Status Tech Risk/Reward
BrainChip BRCHF Public (QN) Neuromorphic edge chips 10
9. FINAL CONSOLIDATED WEIGHTED LIST (RANKED BY RISK/REWARD)
Top Asymmetry (9–10) — “Future Winners if the GPU Era Ends”
1. Groq – 10
2. d-Matrix – 10
3. BrainChip (BRCHF) – 10
4. Cerebras (CBRS) – 9
5. SambaNova – 9
6. Tenstorrent – 9
7. Astera Labs (ALAB) – 9
8. QuickLogic (QUIK) – 9
9. Blaize (BZAI) – 9
10. Lattice (LSCC) – 8 (highest-quality public asymmetry)
Medium Asymmetry (6–8) — “Likely Early Entrants to Replace GPUs in Specific Domains”
1. Rambus (RMBS) – 8
2. SK Hynix – 8
3. Coherent (COHR) – 8
4. Entegris (ENTG) – 8
5. Marvell (MRVL) – 7
6. Broadcom (AVGO) – 7
7. TSMC (TSM) – 7
8. Arm (ARM) – 7
9. Ambarella (AMBA) – 7
10. Micron (MU) – 7
Lower Asymmetry (4–6) — “Solid but mature”
1. NXP (NXPI) – 5
2. Cisco (CSCO) – 5
3. Mobileye (MBLY) – 6
4. Qualcomm (QCOM) – 6
5. UMC (UMC) – 6
6. GlobalFoundries (GFS) – 6
7. Synopsys (SNPS) – 6
8. Cadence (CDNS) – 6
9. Microsoft (MSFT) – 4
10. Amazon (AMZN) – 5
11. Alphabet (GOOGL) – 5