I trace the wallet, not the whisper. When Goldman Sachs projects a 119% profit surge for a hardware supplier in 2028, I don’t reach for my calculator—I reach for my on-chain forensic toolkit. The target here is Zhongji Xuchuang, the Chinese optical module manufacturer riding the AI infrastructure wave. The whisper says 'AI capital expenditure is unstoppable.' The wallet says otherwise.
Context: The Optical Module Kingpin and the Hype Cycle Zhongji Xuchuang is the world’s leading supplier of high-speed optical transceivers—the fiber-optic connectors that link thousands of GPUs inside AI training clusters. Think of it as the pickaxe seller in the AI gold rush. Its 800G modules are already in mass production for NVIDIA, Google, and Microsoft. Now Goldman Sachs has raised its profit forecast by an astonishing 65% for 2026, 108% for 2027, and 119% for 2028, citing 'AI infrastructure expansion' and a shift to 1.6T and 3.2T modules with higher average selling prices. The target price implies a 163% upside from current levels. The bull case is seductive, but as a cold dissector who has audited smart contracts and watched DeFi yield farms implode, I see the structural fragility beneath the surface.
Core: A Systematic Teardown of the Assumptions First, the technology. Goldman’s thesis rests on silicon photonics ramping to volume. I have seen this movie before—in 2021, every NFT project promised AI-generated art, but the code was a simple swap. Silicon photonics has been hyped for a decade. Yield issues in wafer-level optical coupling remain acute. The 1.6T modules require advanced co-packaged optics that most suppliers cannot mass-produce reliably. The analysis I reviewed claims that 'engineering maturity' is sufficient, but my experience auditing the 0x protocol taught me that promises of imminent scaling are often camouflage for unresolved defects.
Second, commercialization. The profit forecast assumes that ASP will rise as technology advances. That is true in the first year of a new generation, but the optical module market is a race to the bottom. Competitors like Coherent, New Yee, and Huawei-backed players are already sampling 1.6T. The moment NVIDIA qualifies a second supplier, pricing power evaporates. I saw this during DeFi Summer 2020—high yields attract capital, which then competes away the yield. The same dynamic applies here: high margins attract competition, which destroys margins.
Third, the customer concentration risk. Zhongji Xuchuang’s top customers are a handful of cloud hyperscalers. Any one of them—Microsoft, Google, or Amazon—could decide to vertically integrate optical connectivity. Microsoft already has its Lyra project for co-packaged optics. If that goes internal, a chunk of Zhongji’s order book disappears instantly. In crypto, we call this a 'rug pull'—except here it’s delivered via an internal memo, not a smart contract.
Contrarian: What the Bulls Got Right To be fair, the bulls correctly identify that AI data centers have a genuine bandwidth bottleneck. The number of GPUs per cluster is doubling every year, and the network fabric must keep pace. Optical modules are not optional; they are the critical path. Goldman’s 800G demand thesis has been validated by actual shipments—Zhongji’s revenue grew 60% year-over-year in Q2 2024. The underlying driver, AI capital expenditure by big tech, is real and likely to persist through 2025-2026. The mistake is extrapolating that linear growth into 2028 without accounting for saturation, competition, and technological substitution.
Takeaway: The Accountability Call When the yield is too high, the exit is rigged. Goldman Sachs predicts a 119% profit surge three years from now. Historical data shows that long-range forecasts by investment banks have a 40% accuracy rate at best. For a hardware company in a cyclical industry, the probability is even lower. The real value of this analysis is not the target price—it is the warning. Blindly following a single source of hype, especially from a blockchain/Web3 media outlet that likely has undisclosed incentives, is the same mistake that led to Terra’s collapse. I trace the wallet, not the whisper. And the wallet here shows an overvalued company with a fragile competitive moat. The lesson for crypto investors is clear: tear down the assumptions, audit the code of capital expenditure, and never mistake a bull market for structural strength.