This week has seen the hype for 5G, the creatively titled successor to 4G mobile internet connectivity, go from a simmer to a boil as Qualcomm, AT&T, Verizon, and Samsung rattled off a series of launch plans for early next year, while Apple was reported to be sitting out the technology until at least 2020. As much as I love the excitement surrounding the introduction of the new cellular networking standard, I’m inclined to think that Apple has it right: none of us should be factoring 5G into our phone purchases in the coming year.

It’s not that 5G isn’t hugely promising. Samsung today issued a press release saying it’s achieved a throughput of 1.7Gbps with a combination of Verizon’s spectrum and Qualcomm’s Snapdragon X50 5G modem. That’s the best-case scenario, but even Qualcomm’s real-world modeling suggests a still-impressive jump to speeds of around 490Mbps, an order of magnitude improvement from the 4G LTE networks we have today that typically offer connections of around 20Mbps to 50Mbps.

But pay attention to the circumstances here: both Samsung and Qualcomm are telling us about their lab results. No carrier or manufacturer is walking around cities with actual 5G devices yet and giving us truly real-world measurements. At Qualcomm’s big 5G get-together in Hawaii yesterday, my colleague Sean Hollister was only shown one Samsung 5G phone, which he wasn’t allowed to touch, a 5G Moto Mod, and a couple of chunky 5G mobile hot spots, and none of those devices exhibited speeds higher than 140Mbps.

Every participant in the limited Hawaii demonstrations had the same boilerplate excuse: the test 5G network was put together in a hurry by Ericsson, and the actual devices and the “tuned and optimized” (in the words of Verizon director of architecture Chris Emmons) US network deployment will indeed deliver multi-gigabit speeds. Will.

Beside still being more a recipe than a fully cooked dish, 5G also has one particularly distasteful aspect. Its fastest speeds come from airwaves in the millimeter-wave frequency range, which has been dubbed mmWave. Verizon has 28GHz, and AT&T operates at 39GHz. Anyone who’s ever tried the failed LTE rival that was WiMAX knows the biggest problem with these high-frequency transmissions: they struggle to penetrate through walls or other physical obstructions. The fallback of lower-frequency, sub-6GHz 5G will happen, we’re told, but “the silicon to support that FDD spectrum doesn’t become available until later in 2019,” according to Gordon Mansfield, AT&T’s VP of converged access and device technology.

To put it simply, the first 5G devices will spend a lot of their time running on 4G LTE networks because that’s the most compatible and available thing they’ll have to connect to.

Even when you do strike a vein of ludicrous-speed 5G in your vicinity, you’ll quickly find that you don’t have all that much you can do with it that 4G cannot already handle. The earlier move from 3G to 4G unlocked some use cases for mobile apps and services — such as storing your photos in the cloud or streaming your music instead of downloading it — that were desperately waiting for more bandwidth. In Hawaii this week, the 5G advocates showed off a streaming virtual reality demo, but VR is yet another piece of technology that’s still evolving and searching for a truly compelling application. Without someone like Nintendo creating an amazing VR game or some other high-bandwidth interactive experience in 2019, 5G’s near-term future will be one of high cost for little reward.

5G in 2019 will face many of the same hurdles that 4G had to overcome in 2011. The new networks will have tiny geographic coverage to begin with, and the silicon required to connect to them will be nowhere near as energy- or space-efficient as the now-mature 4G tech. So the first taste you’ll get of the new hotness in tech will be a rapidly depleting battery on a beefier-than-usual device. (Seriously, there are PCs smaller than that Verizon “puck” above, and AT&T’s 5G hot spot is no smaller.) I’d like to believe the mobile industry is smart enough not to repeat the mistake that was the HTC ThunderBolt — 2011’s poster child for a phone that added 4G too quickly and paid the price in suboptimal battery life and design — but I’m not sure many phone makers will have the freedom to choose.

Samsung, Huawei, and Xiaomi have each signaled their intent to offer a 5G phone in 2019, while OnePlus has even gone so far as to promise that it will have the first 5G phone in Europe. In the hyper-competitive realm of Android smartphones, you either copy Apple gratuitously or you pursue the status of achieving a technical first, whether it’s notchless or foldable displays, in-screen fingerprint scanners, or, now, 5G. Discrete features of this kind are major selling points for manufacturers, so whether the networking or chip technology is there or not, when one company jumps into the fray, all the others inevitably have to follow suit.

Only Apple has shown itself to be (partially) immune to the reckless chase for new features and specs, having previously sat out the 2011 mess of terrible 4G phones before launching its first 4G iPhone the following year. The luxury of Apple’s enduring customer loyalty allows the company to skip the bloodbath of in-development products and just deliver a refined one when the time is right. On this occasion, Apple is also constrained by its relationships: it’s on bad terms with the 5G leader Qualcomm, and its partner Intel won’t have 5G modems ready before 2020. So whether this is Apple being judicious or merely limited by circumstance, its first 5G device is a long way off.

I don’t doubt that 5G will be a massive leap forward in mobile connectivity, to be matched only by our insatiable, irrational hunger for greater speed and bandwidth. But we’ll get only glimpses of that future in 2019, and I don’t think those glimpses will be worth paying for. Over the coming year, carriers like AT&T and Verizon and phone makers like Samsung and Huawei will be figuring out how to optimize their equipment, maximize their coverage, and harmonize the various wireless components that reside under the “5G” banner. Once that work is done, I’ll happily revise my recommendation and urge everyone to board the 5G hype train. Just not yet.