Serendipity35

I wrote last week about plans to harvest solar power from space for places like data centers. If it sounds like science fiction, that might be because it was first imagined in a 1941 short story, "Reason," by Isaac Asimov. (see below)

It was formally proposed by engineer Peter Glasser in 1968, a space pioneer who introduced the idea of using satellites to beam solar energy from space down to Earth. Over the decades, what Glaser envisioned has been known by many names — space-based solar power (SBSP), solar-power satellites or satellite power system (SPS), as well as satellite solar-power system (SSPS). Glaser's contributions to space science and technology were not limited to the solar-power satellite concept. He also worked on NASA's Apollo moon missions and headed an experiment that flew aboard the space shuttle Columbia in 1986.

But solar power from space beamed to Earth has remained mostly theoretical due to cost and complexity.

Data centers need power. A lot of power. People don't want data centers in their neighborhoods. Where will it come from? From space?

Meta announced a deal with startup Overview Energy to purchase solar power collected by satellite and beamed back to Earth.

It is an experimental approach that could power data centers at night. Unlike traditional solar power, which relies on storing daylight, space-based solar power aims to deliver continuous energy.

Overview Energy plans to deploy satellites over 22,000 miles from Earth's equator, where they would collect and transmit infrared energy to solar panels. A test is scheduled for 2028, with a commercial rollout in 2030. Meta is seeking up to 1 gigawatt of power from the project, underscoring its energy needs for AI.

It sounds a bit wishful thinking if you look at the numbers. In 2024, Meta's data centers consumed 18,000 times the electricity that this deal would deliver in a single hour. 

Space-based solar power (SBSP) involves harvesting solar energy in orbit and beaming it to Earth, providing 24/7 clean energy unaffected by weather, nighttime, or atmospheric filtering. There are challanges: high launch costs, complex orbital assembly of massive structures, and wireless energy transfer. 

You may have heard the advice to speakers to open with a joke, so here we go.
A bunch of scientists created a huge machine capable of complex calculations and called it UNIVAC. Eager to test their invention, they asked it, “Is there a God?”The vacuum tubes hummed, and the tape spools spun for several minutes. Finally, the machine spat out a little card, on which was written, “THERE IS NOW.”

That's an old joke, but it seems fresh in this "Intelligence Age" of artificial intelligence and fears of a singularity. In this time of AI and having a computer in the palm of our hand, it is interesting to consider what was happening in tech history back in 1951. That was when the Remington Rand Corporation signed a contract to deliver the first UNIVAC computer to the U.S. Census Bureau.

UNIVAC I (which stands for Universal Automatic Computer) took up 350 square feet of floor space — about the size of a one-car garage — and was the first American commercial computer. It was designed for the rapid and relatively simple arithmetic calculation of numbers needed by businesses, rather than the complex calculations required by the sciences. It was intended to compete against IBM’s punch card-reading computers, but UNIVAC read magnetic tapes, not punch cards, so a special “card to tape converter” had to be designed.

Though the government contract was signed and a ceremony held on March 31, the computer wasn’t actually delivered until the following December. There was only one UNIVAC I, and Remington Rand wanted to use it for demonstration purposes. They asked for and received time to build a second computer.

The government was the first big customer of the UNIVACs, with subsequent models going to the Air Force, the Army Map Service, the Atomic Energy Commission, and the Navy.

The computer first came to the notice of the general public in 1952, when CBS used one to predict the outcome of the presidential election. UNIVAC correctly picked Eisenhower and predicted his electoral count within 1 percent, but the network didn’t release the results until after the election was called, so as not to affect the outcome.

The first commercial sale was to General Electric, for their Appliance Division, followed soon after by the Metropolitan Life Insurance Company, in 1954.

There were 46 UNIVAC I’s built and delivered, in all.

It had been in the news in the United States all week, the federal government is moving to a policy that will require the power-hungry data centers to get out of the public energy pool and go swim in a plasma of their own making.  Big Tech companies are building and investing in their own energy supplies as they race to meet the huge energy demands of AI computing in 20th century datacenters.  It's estimated that a "traditional" (non-AI producing) datacenter rack can consume somewhere between 5 and 15 kilowatts of power --think central A/C units, commercial clothes dryers, banks of EV car chargers.  That same rack, running AI-capable hardware and processing power will consume 10 times as much -- up to around 100 kilowatts.

As robust as the United Stated power grid is, the demand for electricity to power these clustered artificial intelligence entities will exceed the current ability to support that demand. To feed this need for power, Amazon has begun developing small nuclear reactors (SMRs), Oracle and OpenAI are working on half trillion-dollar natural gas fueled electrical plants.  These solutions have their obvious drawbacks:  Amazon's quest for contemporary electricity using the nuclear option will produce the most-toxic waste ever thrown away, and it will last for thousands of years.  Oracle and OpenAI's investment in huge natural gas energy sources risks not only accelerating climate rot, but it risks exhausting energy supplies at scale.  The risks of expanding the electricity supply on a 20th century grid are substantial.  If the demand for electricity was reduced, those risks would subside.

The development of quantum computing has quietly been on the rise.  These computing instances, in total, consume about 25kilowatts for super computers that require extreme refrigeration to drive their super-conductor-based processors at temperatures near absolute zero.  The warmer weather loving neutral-atom computers operate around room temperature and use 7 kilowatts (or less) of power.  When optimization tasks or simulations are sent to their quantum algorithms, these computers produce solutions at orders of magnitude faster and use a tiny fraction of the energy that a traditional datacenter would require.

Quantum computing, now, can significantly enhance AI (Generative AI) by its speed.  Quantum computers are faster and deeper in data analysis and have now led to a new class of Generative AI called GenQAI (Generative Quantum AI) that can use quantum hardware to iterate complex problems and generate more human-like reasoning and intuition in AI.

Quantinuum, which is reported to be one of the world leaders in quantum technology, in November unveiled its Helios system, which has been described as the world's most accurate quantum computer. That quantum instance requires less than 40 kilowatts of power, about the same as a single data center rack average AI Generative load and configuration.  The company announced last week it was going public and would issue an IPO sometime in the first half of 2026.

With some clairvoyant disruption and a little bit of luck, we'll have frugal quantum computing cottages humming their 4-dimensional power song before we have natural gas caverns and poisonous landfill dirges to endure