Applied Optoelectronics Inc. (AOI) will demonstrate a prototype 3GHz amplifier at this week's SCTE TechExpo in Atlanta — the cable industry's first concrete glimpse of HFC network gear that could ultimately deliver 25 Gbit/s downstream speeds.

The demo is deliberately basic. AOI is using a signal generator to produce single-channel QAM and orthogonal frequency-division multiplexing (OFDM) signals running up to 3GHz, feeding them into a proof-of-concept amplifier, and measuring the output with a spectrum analyzer. The goal, according to Rafael Celedon, chief designer for AOI's broadband business, is to prove that generating and amplifying signals at those frequencies is possible and that the demonstration is "real."

But AOI is upfront about the distance between a lab demo and a deployable product. Commercial deployment of 3GHz-based technology on DOCSIS hybrid fiber/coax networks, Celedon said, is years away. "I don't foresee anything that anybody could deploy for years," he told us.

The power problem

The central engineering obstacle is power consumption and the heat it produces. Amplifying signals across the coax portion of an HFC network at 3GHz — rather than the 1.8GHz ceiling of today's most advanced amplifiers — carries severe physical penalties if the current amplifier spacing stays unchanged. Keeping today's amp spacing with 3GHz technology would require roughly 50 times more power and dissipate more than a kilowatt of heat, Celedon explained.

"Those are crazy numbers… Nobody's going to try to do that," he stressed.

The industry's answer, according to AOI, is a more distributed architecture. Rather than pushing a single main amplifier to cover the same span, operators could deploy additional booster amps that reamplify the signal before it reaches the next main amplifier. Celedon said work is underway to develop a way to do this without extensive manual splicing, with one possible insertion point at or near network taps.

"Every tap is an opportunity to boost the signal," said Celedon. "You have to change [the tap] no matter what" when shifting to 3GHz.

That could take the form of a tap/booster combination product that does not yet exist on the market.

Standards groundwork advances

The demo lands as cable technology suppliers, operators and industry bodies push forward on 3GHz specifications. CableLabs has been working for at least a year on an "optional annex" for DOCSIS 4.0 that envisions HFC networks built out to 3GHz, providing the capacity for 25 Gbit/s downstream speeds.

SCTE has already published two ANSI-accredited standards aimed at 3GHz: one for tap housings (ANSI/SCTE 264) and one for passive housings (ANSI/SCTE 265), according to Dean Stoneback, senior director of engineering and standards at SCTE. SCTE will eventually fold CableLabs' work into its standards portfolio and revise existing documents as needed — the same process the organization followed for earlier 1GHz, 1.2GHz and 1.8GHz generations.

Retrofit path, missing chips

One reassuring element for operators: the mechanical framework for 3GHz may largely already be in place. Celedon maintains that current-generation 1.8GHz amp connectors and housings are "mostly ready" for 3GHz, meaning a field-deployed 1.8GHz amp could be retrofitted by swapping certain elements — such as the plug-in module — rather than replacing the entire unit. That approach, however, depends on booster amps being available to handle heat dissipation beyond what a main 3GHz amplifier alone can manage.

The bigger gap is silicon. Celedon identified the chips needed for 3GHz amplifiers as the "main limitation," and those chips do not yet exist. For the TechExpo demonstration, AOI's prototype runs on existing semiconductors from Qorvo and Mini-Circuits, the latter of which acquired Analog Devices' cable amp business in 2024.

AOI has also published a whitepaper laying out potential paths to 3GHz, covering a framework for amplifiers, power, network designs and architectures, and examining questions of affordability, feasibility and whether end customers would notice the difference.

The timeline is instructive. Celedon noted that work on 1.8GHz devices began around 2019 and only largely solidified in 2025. "It takes a while to solve all of these issues," he said. By that precedent, cable operators eyeing 3GHz capacity should expect the technology to mature on a similar multi-year horizon — with the TechExpo demo marking the starting point, not the finish line.