Why Computing Has Stopped Getting Cheaper - And Where the Next Decade of Gains Will Come From
Systemiq Capital's view on advanced computing, the intelligence substrate of the physical economy.
In brief. For fifty years, computing became predictably faster and cheaper as silicon transistors shrank. That shrinking has now stopped paying. The next decade of progress is migrating beyond the transistor itself — into how chips are packaged, how data moves between them, how systems store and retrieve memory, how power reaches the processor, and, in time, into computing that is not conventional silicon at all. Systemiq Capital invests in the companies building this next layer, backing IP-rich, asset-light teams that either bend the demand curve on compute energy or open capabilities conventional silicon cannot reach. The firm has made two investments in the space to date: Mixx Technologies and Claros.
Why has computing stopped getting cheaper?
For half a century, the gains in computing came from shrinking silicon transistors — the trend described by Moore's law. Each generation of smaller transistors made computing faster and cheaper at a predictable rate. That process has now run into the physics of atomic limits, and shrinking the transistor further no longer delivers the cost and performance gains it once did. At the same moment, demand for computing from artificial intelligence has risen sharply.
The result is that the cost and energy curve of computing is uncertain for the first time in fifty years. The gains are migrating away from the transistor and into the rest of the system: how chips are packaged, how data moves between them, how systems remember, how power is delivered, and, in the longer term, into computing paradigms that are not silicon-based at all.
What is the intelligence substrate of the physical economy?
Computing is the intelligence substrate of the physical economy — the layer that turns physical inputs into information, performs computation, and returns a physical output. A sprayer reads a weed in a field and decides to fire. A grid balances supply and demand minute by minute. A model reads a disease target and designs a therapeutic. A satellite decides which of its images matter before transmitting data to the ground.
This substrate is not confined to the data centre or to silicon. It runs from the chip package to the rack to the grid, out to the vehicle, the factory line and increasingly into orbit. It is built from whatever physics does the job best — electrons, photons and, in time, qubits. It also includes the software pressed against the hardware: the kernels, compilers, runtimes and orchestration layers that determine how much useful work the physics actually performs.
For Systemiq Capital, compute sits underneath all three of the firm's investment themes — Electrification, Decoding Nature and Applied AI — which is why advanced computing sits within the Electrification theme and functions as the foundation the whole portfolio stands on.
Why is AI's energy demand a commercial opportunity?
The rise of AI is driving a sharp increase in electricity demand. Advanced economies had seen several decades of essentially flat electricity demand; data centres are now putting the power sector in those economies back on a growth footing, projected to account for more than 20% of electricity demand growth to 2030 [IEA, Energy and AI, 2025]. Global data centre electricity consumption is projected to more than double to around 945 terawatt-hours by 2030 — roughly equivalent to Japan's total electricity consumption today [IEA, Energy and AI, 2025].
The same grids that must electrify transport, industry and buildings are now also feeding computation. Without a step-change in efficiency, AI and electrification will compete for the same electrons. Significant electricity is also lost inside the data centre itself, across successive conversion stages and voltage regulators, before it ever reaches a processor.
The commercial logic is straightforward: efficiency gains in compute are among the highest-leverage opportunities of the decade, and one where the commercial incentive and the energy outcome point in the same direction. Hyperscalers do not buy efficiency because it is virtuous. They buy it because their margins, their capacity to grow and their licence to operate depend on it. Efficiency that customers fight to pay for scales in a way that subsidised change does not.
What does Systemiq Capital back in advanced computing?
Systemiq Capital backs IP-rich, asset-light companies doing one of two things: bending the demand curve on compute energy, or opening capabilities that conventional silicon cannot reach.
The first delivers more useful compute per watt, per rack and per dollar of capital. The second changes what can be computed and where computation can live — from quantum processors that simulate chemistry and materials classical machines cannot, to compute in orbit that puts intelligence where there is no grid and no fibre.
There is no single breakthrough chip coming to reset the curve. The next decade of progress comes from rethinking the entire stack: how electrons enter the data centre, how they are regulated at the chip, how data moves between processors, how heat leaves the rack, and how efficiently the software layer runs intelligence on top. Every layer is now a place where meaningful gains can be won, and every layer has incumbents that struggle to move at startup speed.
Two structural shifts make this newly investable. Hyperscalers now buy and qualify full reference architectures rather than loose components — an OCP-shaped market that rewards small teams able to ship a complete system rather than a single part. And the fabless model, long established in digital logic, has arrived in analog, power and photonics: companies can own the IP at a critical layer while manufacturing partners such as TSMC, Samsung and STMicroelectronics carry production. This removes the balance-sheet requirement that historically kept venture capital out of semiconductors.
What are Systemiq Capital's advanced computing investments?
Systemiq Capital has made two investments in advanced computing to date, both design companies whose manufacturing is carried by leading industry partners, and both founded by teams that had already shipped this class of technology into production.
Mixx Technologies replaces the copper connections inside frontier AI systems with light — the same shift that transformed long-distance telecoms, now arriving inside the rack as co-packaged optics, delivering far more bandwidth per watt.
Claros designs a voltage regulator small enough to sit directly beneath the AI chip, so that power arrives precisely where it is needed rather than being lost as heat along the way.
The two bookend the rack: Mixx addresses how data moves through it, Claros how power enters it.
Where is Systemiq Capital looking next in compute?
Four themes are shaping the firm's pipeline.
Inference is moving to the edge. Training frontier models will stay in hyperscale data centres, but using them increasingly happens where the data is — on the robot arm, in the vehicle, at the factory line, on the satellite. Latency matters when a machine makes a physical decision; sensitive data is better processed where it is generated; and every watt of inference handled at the edge is a watt the grid never has to deliver to a data centre. The opportunity is in inference-first silicon, low-power accelerators, and the quantisation, sparsity and compilation techniques that let capable models run within tight energy budgets. Portfolio company Archetype AI runs its physical agents on machines themselves rather than in the cloud, because the major cloud models cannot be adapted to do so.
Memory is becoming the bill. The industry conversation has moved from training models to serving them, and the bottleneck has moved with it. Serving a model is two jobs — reading the prompt, which is hungry for raw compute, and writing the answer, which is hungry for memory — and leading operators are beginning to split the two across different hardware. On long tasks, the working memory a model holds (the KV cache) can outgrow the model itself. Increasingly, the cost of an AI answer is set not by how fast the chip thinks but by how much the system can remember and how quickly it can fetch it. The response is a new memory hierarchy, tiering from HBM on the GPU down through CXL-attached system memory and flash, alongside hardware that brings computation closer to where data lives. Systemiq Capital is actively looking for companies here.
Packaging and connectivity are converging. With the transistor no longer shrinking cheaply, performance gains have moved to how dies are stacked, bonded and connected — advanced packaging, which is really scaling inwards rather than downwards. Packaging capacity, not chip design, is now among the tightest constraints on AI hardware supply. The toolkit is expanding fast: logic stacked on logic, memory bonded onto compute, copper-to-copper hybrid bonds replacing solder bumps, and interposers evolving from passive layers into active ones that route power and signal themselves. At the same time, the optical boundary is crossing into the package itself: when Systemiq Capital invested in Mixx, co-packaged optics was a contrarian position; it has since entered production at the industry's largest players, with UCIe and the first co-packaged optics standards emerging. The open question is who supplies that ecosystem — the lasers, the fibre attach, the connectors, the test and reliability infrastructure.
Power electronics have found their own Moore's law. While transistor scaling stalled, a quieter improvement curve kept compounding. New semiconductor materials — silicon carbide and gallium nitride — handle high voltages and heat with a fraction of the losses of ordinary silicon, and each generation gets denser and more efficient. That curve made electric vehicles viable and is now transforming how data centres deliver power, with the move to 800V DC distribution running from the grid connection down to regulators millimetres from the chip. Each stage of that descent — from tens of kilovolts at the fence line to under a volt at the transistor — is its own conversion problem, and its own opportunity. The next frontier is space, where every gram and every watt is priced brutally, and where power infrastructure for satellite constellations and in-orbit compute must be dense and fault-tolerant from day one.
Where do the companies come from?
The United States has the greatest density in advanced computing, because the customers and strategic acquirers are concentrated there and proximity compounds. Both of Systemiq Capital's first two investments are American companies. Europe holds genuine pockets of strength: AI research in London, semiconductor tooling in the Netherlands, robotics in Zurich and more, with the best European teams now building global companies from day one. As a transatlantic investor, Systemiq Capital's role is to be the bridge — helping European founders reach the ecosystem where their customers sit, and bringing an independent perspective to American companies from outside the Valley.
What kind of founders does Systemiq Capital back in compute?
The founders who win in advanced computing understand that it is a game of system design, manufacturing readiness and supply chain relationships. Some learned this over decades inside the industry's largest companies; others are younger builders who surround themselves with people who did. What they share is a conviction that reliability and manufacturability beat benchmarks, and that the supply chain is a first-class part of the product.
Frequently asked questions
What is the physical economy? The physical economy is how energy is traded and managed, and how natural and industrial systems become more intelligent, productive and resilient. Systemiq Capital backs founders redefining it across three themes: Electrification, Decoding Nature and Applied AI.
Why does Systemiq Capital invest in advanced computing? Computing is the intelligence substrate underneath all three of the firm's investment themes. As the cost and energy curve of computing becomes uncertain for the first time in fifty years, the companies rebuilding the compute stack — in packaging, power, memory and connectivity — are central to the systems Systemiq Capital invests in and to the energy transition itself.
What does Systemiq Capital look for in a compute company? IP-rich, asset-light companies that either bend the demand curve on compute energy (more useful compute per watt, per rack and per dollar) or open capabilities conventional silicon cannot reach. Manufacturing is typically carried by industry partners, and the strongest teams treat manufacturability and the supply chain as core to the product.
What stage and geography does Systemiq Capital invest at? Series A +/– across the UK, Europe and the US.
Which compute companies has Systemiq Capital backed? Mixx Technologies, which replaces copper interconnects inside AI systems with co-packaged optics, and Claros, which designs a voltage regulator that sits directly beneath the AI chip. [FLAG: confirm public before publishing.]
Written by Jasper Wigley, Principal, Compute, Systemiq Capital.